747
J. Org. Chem. 1993,58, 747-749
Characterization of the Bisketene Photoisomer of Benzocyclobutenedione Thomas Mosandl and Curt Wentrup' Department of Chemistry, The University of Queensland, Briebane, Queensland 4072, Australia Received June 29, 1992
The photochemistry of benzocyclobutenedione (1) has been the subject of several detailed investigations under solution conditions,l12as a thin f i i 3 and in glassy matrices4 at -196 "C, in Ar matrices,596 and under laser flash conditionsat room tem~erature.~ Nevertheless, there are still problems with the primary photochemistry. It has been reported that the photolysis of 1 produces two intermediates, assigned as bisketene 2 and carbene 6,1323-7and that under matrix conditions benzyne (3) is ultimately
-
hv
hv
a
1
3
9
0
8
4
I
0
8. (39%)
ylide formation with pyridine and adamantanethione and trapping with methan01.~ It should be noted that there are several examples of rapid reaction between ketenes and pyridine or other tertiary amines, giving zwitterionic intermediates at room temperat~re.~
Experimental Section Benzocyclobutenedione (1),lkmethyl 2-formylbenzoate@),'Ob and 3-methoxyphthalide (6b)' were prepared according to literature methods. Dimethyl phthalate (4) was obtained from Aldrich Chemical Co. Ar matrix isolation was performed on a Leybold-Heraeus closed-cycle liquid He cryostat at 11-25 K (2 X mbar)." Photolyses were carried out using a Hanovia high-pressure XeHg lamp (lo00 W)equipped with a monochromator. FT-IR spectra were recorded on a Perkin-Elmer 1720X spectrometer. GC-MS was recorded on a Hewlett-PackardModel 5970 MSDI 5890 GC using a 25-m (i.d. 0.22 mm) BP5 capillary column at an initial temperature of 100 "C, programmed to 270 O C at a rate of 16 "C/min;injector temperature200 "C;column head pressure 135 kPa. Retention times: 1,5292 min; 8,5.763 min; 6b,6.240 min; 4,6.658 min (base-line resolution; mean half height peak width 0.03-0.04 min). Trapping of Ketene 2 with Methanol. Ketene 2 was generated by photolysis of 1 at X = 290 nm at 11K in an ArMeOH (-1ml) matrix. After ca. 75% conversion of 1 to 2 (followed by IR spectroscopy), the Ar was removed by gentle warming to 80 K. The system was then flushed with Ar gas and allowed to warm to room temperature. The material on the sample target was examined by GC-MS as detailed above. Only four compounds were detected. They were identified by direct GC-MS comparison with the authentic materials and found to be (with relative peak areas) the following: 1 (1.16),8 (2.30), 6b (2.71), and 4 (1.00).
Results and Discussion
After direct observationof bisketene2 in an Ar/methanol matrix, it has been trapped with methanol to givedimethyl phthalate 4.5 This is somewhat surprising since the reaction2 4 constitutesan oxidationreactionwith formal loss of H2. One may understand such oxidations taking place in solution photochemistry,8 but it is not obvious that 4 should be the only product under matrix conditions. Bisketene 2 has been observeddirectly by IR spectroscopy and was first reported to absorb at 2091 cm-l in an Ar matrix;6 more recently, it was claimed not to absorb at 2091 but instead at 2064 and 2125 cm-l.6 The bisketene 2 has also been trapped in [2 + 41 cycloadditionreactionsin solution,and carbene 5 has been trapped in [2 + 21 cycloadditions to alkenes and alkynes under such conditions.2 A mixture of dimers derived from both 2 and 5 also results from solution photolysis of 1 . l ~ ~ In the laser flash photolysis study? 2 was identified as a long-lived intermediate ( T > 100 ps). Carbene 5 has not been observed directly, but its existence was inferred from
Photolysis of 1, isolated in Ar matrix at 11 K, at X = 290 nm produced ketene 2 on the basis of the evidence presented below. The new species absorbed at 2138,2077, 1595,1346,1080, and 965cm-l and thus had no frequencies in common with those reported earlier.5*6The intensity ratio of the two bands at 2138 and 2077 cm-l is ca. 1:1,and they are by far the strongest bands in the spectrum (Figure 1). They are ascribed to the coupled vibrations of the two ketene functions in 2. The reaction 1 2 is photoreversible, as previously stated by Chapman5without any details. Thus, photolysis of 2 at X > 320 nm regenerates 1. One cycle (1 2 1) is illustrated in Figure 1. It can be repeated several times before 2 is finally converted to benzyne (3). A t 290 nm there was ca. 80% conversion of 1 to 2 in 1 h, and no benzyne was detectable. After 2 h, there was 100% conversion, and traces of benzyne were now detectable (vide infra). Using X = 280 nm, the rate of formation of 2 was about the same, but benzyne formation
(1) Brown, R. F. C.; Solly, R. K. Tetrahedron Lett. 1966, 169-174. (2) (a) Staab, H. A.; Ipaktschi, J. Tetrahedron Lett. 1966, 583-589. (b) Staab,H. A.; Ipaktschi, J. Chem. Ber. 1968,101,1457-1472. ( 3 )Chapman, 0. L.; McIntosh, C. L.; Barber, L. L. J. Chem. SOC., Chem. Commun. 1971, 1162-1163. (4) Kolc, J. Tetrahedron Lett. 1972, 5321-5324. (5) Chapman, 0. L.; Mattes, K.; McIntosh, C. L.; Pacansky, J.; Calder, G. V.; On,G. J. Am. Chem. SOC.1973,95,6134-6135. (6) Simon, J. G. G.; Mllnzel, N.; Schweig, A. Chem. Phys. Lett. 1990, 170, 187-192. (7) Boate, D. R.; Johnston, L. J.; Kwong, P. C.; Lee-Ruff, E.; Scaiano, J. C. J. Am. Chem. SOC.1990,112,8868-8863.
(9) (a) Moore, H. W.; Duncan, W. G. J. Org. Chem. 1973,38,156. (b) Satchell,D. P. N.; Satchell,R. S . Chem.Soc.Rev. 1975,4,231. (c) Seikaly, H.R.;Tidwell,T. T. Tetrahedron 1986,42,2587. (d)Wynberg,H.;Staring, E. G. J. J. Am. Chem. SOC.1982,104.166. (e) Barra, M.; Fisher, T. A.; Cernigliaro, G. J.; Sinta, R.; Scaiano, J. C. J. Am. Chem. Soc. 1992,114, 2630. (f)Kappe, C. 0.; FHrber,G.; Wentrup, C.; Kollenz, G. J. Org. Chem., in preee. It is shown in thie paper that other dipoplar speciee such as dimethyl sulfoxide and triphenylphosphme oxide act in a way similar to pyridine, catalyzing ketene dimerization, presumably via intermediate zwitterion formation. (IO) (a) Cava, M. P.; Napier, D. R.J. Am. Chem. SOC.1957,3606-3607. (b) Brown, C.; Sargent, M. V. J . Chem. SOC.C 1969,1818-1820. Eliel, E. L.; Burgstahler, A. W. J. Am. Chem. SOC.1949, 71, 2251. (11) Wentrup, C.; Blanch, R.; Briehl, H.; Gross, G. J. Am. Chem. Soc.
-
(8)Small amounts of phthalaldehyde and phthalic acid were obtained on the photolysis of ninhydrin in aqueoue solution: Matauura, T.; Sugae, R.; Nakashima, R.; Omura, K. Tetrahedron 1968, 24,6149-6156.
--
1988,110, 1874-1880.
0022-326319311958-0747$04.00/0 Q 1993 American Chemical Society
-
748 J. Org. Chem., Vol. 58, No.3, 1993
Notes
A
I
I
I
23 02200
20ob.o
1900
I
I
1800
1600
1700
I
1500
I
1400
I
I
1300
1200
I
1100
I
1000
I
900
B(
.o
CU-1
Figure 1. Difference-FTIRspectra of the photolyses of 1 and 2. Upper trace: conversion of 1 (negative peaks) to 2 (positive peaks) at X = 290 nm. Lower trace: conversion of 2 (negative peaks) to 1 (positive peaks) at A > 320 nm. Both spectra in Ar matrix at 11 K. Bands due to 2 are at 2138,2077,1696,1346,1080, and 966 cm-l.
a
1
"
I+
li = 280-300 m,
2co
1054,1038,848,736, and 469 cm-l. The very weak triple bond stretching vibration near 1846 cm-1 was not observed.l3 Difference spectroscopy clearly demonstrated that the IR bands due to 2 disappeared as those of 3 and CO (2140 cm-l) formed (290 nm; 10 h).14 We have also found that the bisketene 2 thermally reverts to 1:
3 120 -140 K
was now much faster; it was complete in 4 h. The production of 2 was slower at X = 300 nm (full conversion of 1to 2 in 5 h). Prolonged irradiation at X = 300 nm ( 6 7 h) afforded new species absorbing at 2086 and 2066 cm-l, together with benzyne. The species absorbing at 2086 cm-1 is presumably cyclopentadienylideneketene (7), obc=c=o 7
served by Brown et d.l2at 2089 cm-l, by Schweig et al.* at 2087 cm-l, by Radziszewski et al.13 at 2086 cm-l (Ne matrix), and by us at 2090 cm-l (Ar,12 K)or 2080 cm-' (neat)." Prolongedirradiation of 2 leads tobenzyne(3)(complete in 10 h at 290 nm) and CO. No other producta or intermediates were observed at this wavelength. In agreement with literature data5919 benzyne absorbed at (12)Brown, R.F.C.; Browne, N.R.; Codeton, K. J.; Eastward, F. W.; Irvine, M.J.; Pullin, A. D. E.; Wienum, U. E. A u t . J. Chem. 1989,42, 1321-1344. (13)Rndzhmki,J. G.;H m , A. B.;Zahradnik, R.J. Am. Chem. SOC. 1092,114,52-67.
a
1
It was possible to generate neat 2 by photolysis of a thin (14)(a) There hm been nome confusionin the literature over the CIC stretching vibration of benzyne, which wm originally reported at 2086 cm-1 (Armatrir),14b a value that hm been repeatedly stated by others.1" However, it is now knownthat the cd:stretch appeare at 1848cm-1 (Ne matrix)13 in agreement with the gm-phtue value of 1860 cm-1 derived from the photoelectron spectrum of C&-.16 We fmt suggested that a weak band at 2080 cm-1 (neat in phthalic anhydride) in the product of t h h vacuum pyrolyab ( F W )of phthalic anhydride could be due to cyclopentadienylideneketane(7),andwewereabletoehowthathatthieepeciee persisted at tamperaturea where bencyne hm disappeared, even though we were unable at that time to excludethat benzyne, too,had an aboorptioa near 2080 cm-l.ll We have repeated and codmed theae obrvationa regarding the FVP of phthalic anhydride. We ab0 agree with other authors6J3JSthat benzyne d w not ~OEEWOa band near 2080-2090 cm-* in the IR since we can w i l y generate I ) without any 7 b e i i preaent. (b) Chapman, 0. L.;C u , C.-C.; Kolc, J.; Rorrenquist, N.R.; Tomioh, H. J. Am. Chem. SOC.1976,97,6686-6688. (c) Dunkin, I. R.;MscDonsld, J. G. J. Chem. SOC.,Chem. Commun. 1979,772. Nam, H.-H.; Leroi, G. E. J. Mol. S t r u t . 1987,157,301-304. &, J. W.; Berry,R S. J. Am. Chem. SOC.1976,618,660. Nam, H.-H.; Leroi, 0.E. Spectrochim. Acta, Part A 1986,41,67-73. Nam, H.-H.;Leroy,G.E. J. Am. Chem.Soc. 1988, 110,4096-4097. Schweig, A,, MQnzel,N.;Meyer, H.; Heidemich, A. Struct. Chem. l989,1,89-100. (15)Leopold, D.G.;Miller, A. E. 5.;Lineberger, W. C. J. Am. Chem. SOC.1986,108,1379-1384.
J. Org. Chem., Vol. 58, No.3, 1993 749
Notes
film of 1 at 11 K. Under these conditions, 1 absorbs at 1768,1787,and 1816cm-l in the IR and 2 at 2076 and 2135 cm-l. Warming of the film containing 2 to 120-140 K caused conversion of 2 to 1, as seen by difference spectroscopy. For further proof of the identity of ketene 2, it was generated, observed, and trapped in an argon-methanol matrix (ca. 1OO:l) by photolysis of 1at 290 nm. After ca. 75 % conversion of 1,the system was warmed, allowing Ar to evaporate, and then flushed with Ar until room temperature was reached. The products isolated at room temperature were examined by GC-MS, which revealed only three compounds, 8,6b, and 4, apart from residual starting material (1)(SeeExperimentalSectionfor details).
into the 0-H bond of methanol. However, under our reaction conditions, a direct addition of methanol to bisketene 2 is more likely, taking place via intermediates 9 or 10 and 11: CO
Wi \
2
-
I A-
u
HQCH3
1
I
i i 0
0
-
HOCH3 ..........
OCH, _.
6 U
2
8
9
P
8b
4
Dimethyl phthalate (4) is indeed a minor product. It corresponds to oxidation of a nonobserved dihydrophthalate. The important product 8 corresponds to addition of methanol to 2:
a
8
V
The other major product, 3-methoxyphthalide (6b), could conceivably be formed from carbene 5 by insertion (16) Andraos, J.; Kresge, A. J.; Peterson, M. R.; Czismadia, I. G. J. Mol. Struct. 1991,232,165-177. Allen, A. D.; Andraos, J.; Kresge, A. J.; McAlliiter, M. A.; Tidwell, T. T. J. Am. Chem. SOC. 1992, 114, 1878. Urwyler, B.;Win, J. Angew. Chem., Znt. Ed. Engl. 1990,29, 790-792. (17) (a) It may be of interest to note that the thermal chemietry of 1 and other precursors of benzyne (3) is in nead of revison 88 regards the claimed observation of PE spectra of 3.17b@We show elsewhere that the
ionization potential ~ l a i m e d ~for ' ~benzyne .~ near 9.23 (9.24)eV is not due to benzyne. It is, in fact, due to the stable compound benzene!17d (b) Dewar, M. J. S.;Tien, T.-P. J . Chem. SOC., Chem. Commun. 1985,12431244. (c) Schulz, R.;Schweig, A. In Structure and Reactivity; Liebman, J. F.,Greenberg,A.,Ede.;VCHNew York, 1988,pp284-367 (inparticular pp 334-336). (d) Senio, A.; Gracian, F.; Pfister-Guillouzo,G.; Mosandl, T.; Wentrup, C. Unpublished results.
8b
11
The hydration of ketenes to give enediols (enols of carboxylic acids) has been thoroughly studied in recent years and provides a good analogyfor the enol intermediate 10.16 We have no direct evidence for the presence of 5 in our matrices where it would be expected to exhibit a strong C 4 band above 1750 cm-l. (There were weak bands near 1850 cm-l in several matrices, in particular in the Ar-MeOH matrix, but this is not an ideal medium for IR spectroscopy; the 1850 cm-' species may be due to benzocycl~propenone.~~) Therefore, we prefer to rationalize the formation of the products (8,6b, and 4) in terms of the observable intermediate 2." In conclusion, 5,6-dicarbonylcyclohexa-1,3-diene(2), formed by photolysis of benzocyclobutenedione (l),absorbs in an Ar matrix at 2077 and 2138 cm-'. 2 thermally reverts to 1 at 120-140 K and photochemically at X > 320 nm. The reaction 1 e 2 can be driven in both directions severaltimes before significantdecarbonylationtobenzyne occurs. In agreement with the observations of others, benzyne does not absorb in the IR near 2080-2090 cm-l. Trapping of bisketene 2 in an Ar-MeOH matrix leads to methyl 2-formylbenzoate (81, 3-methoxyphthalide (6b), and dimethyl phthalate (4) (2.3:2.7:1 by GC-MS). The carbene 5 was not directly observable in this work, and the isolation of 3-alkoxyphthalides(6) from the photolysis of 1in alcoholsshould not be taken as unequivocalevidence for the formation of 5 in these reactions. Acknowledgment. This work was supported by the Australian Research Council. T.M. thanks the Deutache Forschungsgemeinschaft for a postdoctoral fellowship.
